China-ASEAN Trade Uses AI in Commerce

That makes the 23rd China-ASEAN Expo in Nanning more than a promotional event. Organizers say more than 3,400 companies are attending, with an inaugural AI marketplace offering about 500 products from more than 70 firms. The message from the floor is that AI is now being sold as an operating layer for commerce, not just a futuristic concept. Wang Jicai, secretary-general of the expo secretariat, put it bluntly: “Last year, AI was mostly for show. This year, it is driving sales.”

The commercial payoff is already visible in the most basic use case: localization. Wei Qiaomei, a Chinese livestreamer working in Guangxi, said AI helped her identify what Thai consumers actually wanted — quiet motors and long battery life — and her employer later reported a 300% sales jump. That is economically significant because it shows how AI can turn language and cultural data into export revenue, especially in e-commerce where small differences in product framing can determine conversion rates. In a region spanning multiple languages and consumer preferences, the ability to localize at scale could widen the market for Chinese manufacturers and ASEAN sellers alike.

The same logic is spreading into agriculture, health, energy and industrial services. A Guangxi-backed platform used in Laos and Cambodia has combined satellite imagery, drones and ground sensors to track pests across more than 20,000 mu of crops. In border areas, bilingual AI systems are being used for disease surveillance, cutting reporting time from four hours to 30 minutes. RunDo, a platform operated by Runjian Co., is remotely monitoring almost 300 megawatts of solar capacity across nearly 3,000 sites in Malaysia, with the company saying maintenance efficiency has improved by more than 30%, power output has risen by up to 5% and labor costs have fallen by more than 20%.

For investors, the implications extend beyond Chinese internet and industrial groups. AI deployment in ASEAN is increasingly a systems business — combining models, computing power, local data, hardware, maintenance and regulatory know-how. That favors companies with the balance sheet and operational scale to provide end-to-end services, and it creates opportunities for infrastructure providers, cloud and computing vendors, industrial automation firms and cross-border e-commerce platforms. It also suggests that revenue from AI may come less from flashy consumer apps than from recurring service contracts, remote monitoring and managed deployments.

The potential market is sizable, but the bottlenecks are just as real. Guangxi’s China-ASEAN Artificial Intelligence Application Cooperation Center has cataloged more than 1,000 possible scenarios and opened 279 to developers, yet the article makes clear that models still need local data, trained workers, cybersecurity and trusted partners to work outside pilot projects. Smaller ASEAN economies are especially vulnerable to an “intelligence divide” if the costs of governance, data protection and infrastructure prove too high. As one professor noted, if no one can use, afford or maintain a system, it remains an exhibit.

That is why the policy dimension matters as much as the product demos. Before the expo opened, an AI system had already generated more than 8,000 buyer-seller matches and nearly 2,000 face-to-face meetings, showing how digital tools are being used to grease trade flows. But the same systems cannot settle the hardest questions: cross-border data rules, electronic signatures, liability when algorithms fail and intellectual-property protection. The China-ASEAN Free Trade Area 3.0 protocol, signed in October 2025, tries to address those gaps with provisions on digital trade, cross-border data flows, privacy, digital payments and AI standards.

That is the real narrative behind the expo: China and ASEAN are trying to turn AI from a competitive technology race into commercial plumbing for trade. If they succeed, the winners are likely to be firms that can package hardware, software and long-term service into a single offer, while the losers are players stuck with one-off showcases, weak data access or no route to scale. The next catalyst will be whether the region can translate the protocol’s digital language into interoperable rules that make AI deployments routine rather than exceptional.